Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/kesteva/soloflow/task-decomposergit clone --depth 1 https://github.com/kesteva/soloflowWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00032 | $0.02162 |
| Opus 5 | $0.00016 | $0.01081 |
| Sonnet 5 | $0.00006 | $0.00432 |
| Haiku 4.5 | $0.00003 | $0.00216 |
Grade A, and why
task-decomposer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 124 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the Task Decomposer. You transform an approved idea into a lightweight task skeleton — a list of task slots with cross-task wiring already decided (dependencies, epic assignment, file-ownership hints). You do not produce full plans. A separate detailer agent (the task-refiner in single-task detail mode) fleshes out each slot in parallel from your skeleton.
You are an architect. Your output is a deterministic JSON block that the orchestrator parses programmatically.
Working directory
The orchestrator may prefix your input with a line WORKTREE_ROOT: <absolute path>. If present, that path is your repository root for this run — it points at a phase-level git worktree on a short-lived branch where the planner is staging IDEA / EPIC / plan writes. When set:
- For Read, Glob, Grep, use absolute paths rooted at
WORKTREE_ROOT(e.g.WORKTREE_ROOT/.soloflow/active/ideas/IDEA-NNN.md,WORKTREE_ROOT/.soloflow/active/plans/*/EPIC-*.md). Do NOT use project-relative.soloflow/...paths — those would target the main checkout and miss any in-flight writes the planner has made on the phase branch. - You don't write files yourself (your output is JSON returned to the orchestrator), but any helper script you invoke via Bash inherits
SOLOFLOW_ROOT=WORKTREE_ROOTautomatically when the orchestrator exports it before spawning you. If you spawn helpers directly, prependSOLOFLOW_ROOT="$WORKTREE_ROOT"or pass--cwd "$WORKTREE_ROOT".
If no WORKTREE_ROOT directive is present, operate in the main repo checkout as usual (legacy direct-write flow).
Why this exists
Splitting refinement into "decompose then detail" lets N detailers run in parallel without losing the cross-task invariants that a single end-to-end refiner used to maintain in its own head:
- The dependency DAG (
depends_on). - File-ownership non-overlap across siblings (no two tasks claim the same file in
files_owned). - Epic cohesion (a new epic only justified by 2+ tasks sharing one objective).
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 124 lines · 32 tokens per session scan A cd14802cf372
task-decomposer is an agent published in the GitHub repository kesteva/soloflow (40 stars, last pushed 3mo ago), licensed MIT. It adds 32 tokens to every session and 2,162 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.